Biology subjects

Fan, X.

Publications and source records attributed to Fan, X..

At least 55 records · Page 3Linked to original sources

bioRxiv2025

Interdependent regulation of trabecular meshwork cell physiology and intraocular pressure by KALRN and TMCO1

Glaucoma is a leading cause of irreversible blindness. Primary open-angle glaucoma (POAG) is its most common form. Higher intraocular pressure (IOP), resulting from impaired aqueous humour outflow, is the cardinal mediating factor, and all proven treatments aim to lower IOP. POAG is a complex genetic disease with numerous loci linked to POAG and higher IOP. The mechanisms by which risk alleles cause disease remain unclear. Here, using primary trabecular meshwork (TM) cells, a cell type controlling outflow, and mice, we found that the POAG-associated gene KALRN encodes an endoplasmic reticulum (ER)-associated Rac regulator essential for TM homeostasis and normal IOP. KALRN loss caused widespread disruption from ER to calcium homeostasis and energy metabolism, leading to induction of cell senescence. KALRN-depletion also led to suppression of the ER translocase and calcium regulator TMCO1, encoded by a gene at one of the most significantly associated genomic loci for POAG. TMCO1-depletion in vitro and in vivo phenocopied KALRN-induced phenotypes, and reduced KALRN expression. These findings establish KALRN and TMCO1 as interdependent regulators of TM homeostasis and IOP, that link regulation of ER and intracellular calcium homeostasis to cell and tissue physiology, and illustrate how different genes linked to glaucoma can form regulatory pathways.

cell biology↗

bioRxiv2025

Inhibition of autophagy-lysosomal function exacerbates microglial and monocyte lipid metabolism reprograming and dysfunction after brain injury

CNS has an overall higher level of lipids than all tissues except adipose and contains up to 25% of total body cholesterol. Recent data demonstrate a complex crosstalk between lipid metabolism and inflammation, suggesting potential contribution of the lipid-rich brain environment to neuroinflammation. While recent data support the importance of brain lipid environment to inflammatory changes observed in age related chronic neurodegenerative diseases, in vivo interactions between lipid environment, lipid metabolism and neuroinflammation in acute brain disease and injury remain poorly understood. Here we utilize a mouse model of traumatic brain injury (TBI) to demonstrate that acute neurotrauma leads to widespread lipid metabolism reprograming in all microglial and brain associated and infiltrating monocyte populations. Additionally, we identify unique microglial and monocyte populations with higher degree of lipid metabolism reprograming and pronounced accumulation of neutral storage lipids, including cholesteryl esters and triglycerides. These lipids accumulate not only in lipid droplets but also in the microglial and monocyte lysosomes and are associated with lysosomal dysfunction and inhibition of autophagy after TBI. Our data indicate that lipid accumulation in these cells is the result of altered lipid handling rather than lipid synthesis and is triggered by phagocytosis of lipid-rich myelin debris generated after TBI. Finally, we use mice with autophagy defects in microglia and monocytes to demonstrate that further inhibition of autophagy leads to more pronounced lipid metabolism reprograming and exacerbated cellular lipid accumulation. Our data suggest a pathological feedback loop, where lipid phagocytosis causes inhibition of autophagy-lysosomal function, which in turn exacerbates cellular lipid retention, reprograming and inflammation.

neuroscience↗

bioRxiv2025

scYeast: a Biological-knowledge-guided Foundation Model on Yeast Single-Cell Transcriptomics

Though large-scale pre-trained models are vital for foundational cell modeling, most focus on human or mouse systems, neglecting model organisms like yeast (Saccharomyces cerevisiae) and often failing to use existing biological prior knowledge effectively. To address this, we present scYeast, the first foundational cell model tailored for yeast that seamlessly embeds biological priors. scYeast employs a novel asymmetric parallel architecture to infuse transcriptional regulatory information directly into the Transformers attention mechanism, leveraging established biological knowledge during training. Pre-trained on large-scale yeast single-cell transcriptomics data, scYeast demonstrates strong generalization and biological interpretability. It excels in zero-shot tasks, such as inferring regulatory relationships and identifying critical cell states. After fine-tuning, scYeast performs exceptionally across a diverse set of tasks from cell type classification to predicting growth doubling time and gene perturbation response. Additionally, using transfer learning, scYeast can be adapted to other omics datasets, such as proteomics, thus broadening its utility in big data analysis. Overall, scYeast is a powerful tool for yeast single-cell biology research and sets a new standard for integrating foundational models with biological prior knowledge, dramatically accelerating the pace of discovery in yeast synthetic and systems biology and providing a replicable framework for other organisms.

bioinformatics↗

bioRxiv2025

Light-Activated Nucleic Acid Amplification Systems Using Photo-caged DNA Polymerase or Primers

Precise control over nucleic acid amplification is essential for achieving reliable, quantitative, high-throughput, and multiplexed molecular diagnostics, particularly in point-of-care and field settings. However, conventional isothermal amplification methods, especially rapid assays at ambient temperature such as recombinase-aided amplification (RAA), suffer from spontaneous initiation and lack of synchronization across parallel reactions. Here, we report light-activatable RAA systems employing two distinct photocaging strategies to achieve spatiotemporal control. In the first, a photocaged DNA polymerase was engineered via site-specific incorporation of p-azido-L-phenylalanine and conjugation with a 2-nitrobenzyl-modified ssDNA blocker through click chemistry. This approach offers high stability, making it advantageous for long-term storage and lyophilized formats. In the second, 5'-nitrobenzyl-modified primers were synthesized to block hybridization or extension until UV activation, offering operational simplicity and rapid integration into existing workflows. Both strategies independently suppressed background activity in the dark and restored full amplification efficiency upon 365 nm UV exposure. The systems exhibited high sensitivity, low background, and compatibility with portable diagnostics. A custom-built device integrating UV activation and real-time fluorescence detection enabled seamless, on-demand operation. This modular platform provides flexible light-gated amplification solutions, allowing the selection of polymerase- or primer-based caging according to storage stability or workflow requirements, thereby advancing precise and field-ready molecular diagnostics.

molecular biology↗

bioRxiv2025

Efficient Cell Factory Design by Combining Meta-Heuristic Algorithm with Enzyme Constrained Metabolic Models

The rational design of high-performance microbial cell factories remains a central challenge in sustainable biomanufacturing due to the complexity of metabolic networks and the difficulty of predicting synergistic genetic interventions. Despite recent advances in strain design algorithms, predicting combinatorial targets remains computationally prohibitive due to the combinatorial explosion. Here, we present MetaStrain, a unified computational framework that integrates enzyme-constrained models (ecModels) with meta-heuristic algorithms to identify non-intuitive combinatorial gene targets for improving product yields. MetaStrain first performs pre-screening through a modified enforced objective flux scanning algorithm to reduce the dimensionality of candidate genes and annotate editing strategies. The subsequent search module employs flexible individual encodings compatible with diverse meta-heuristic algorithms and adopts a bottom-up phenotype evaluation strategy enabling efficient exploration of the combinatorial design space. Integrated redundancy analysis tools further identify single and fixed-size combinatorial strategies, facilitating direct experimental implementation. Computational simulation in Saccharomyces cerevisiae reveals significant enhancements in 2-phenylethanol and spermidine biosynthesis, while controlling target count and covering experimentally validated targets. Experimental validation in Escherichia coli further confirmed the algorithms predictive power, achieving up to a 61.25% increase in L-tryptophan titer of the five-target combination strain. Overall, MetaStrain achieves high computational efficiency, stable convergence, and broad adaptability across diverse metabolic targets. This study provides a powerful tool for metabolic engineering, bridging computational prediction and experimental realization, and highlighting the potential of meta-heuristic optimization in synthetic biology.

bioinformatics↗

bioRxiv2025

Single-nucleus RNA sequencing reveals the cellular diversity of cerebrospinal fluid in the context of intraventricular hemorrhage

BackgroundIntraventricular hemorrhage (IVH) is a common and severe complication of hemorrhagic brain injury. Current treatments offer limited improvement in long-term neurological outcomes. Inflammatory responses in the cerebrospinal fluid (CSF) after IVH are thought to drive secondary injury, but the cellular mechanisms underlying this inflammation remain poorly defined. MethodsWe performed single-nucleus RNA sequencing of leukocytes isolated from CSF collected through external ventricular drains in subjects with intracerebral (n = 6) or subarachnoid (n = 1) hemorrhage. We characterized transcriptionally distinct subpopulations of neutrophils, monocytes, and lymphocytes by comparison to reference datasets. Cell-cell signaling networks were analyzed to infer cytokine-mediated communication, and a flow cytometry panel was developed to validate transcriptomic findings in independent CSF samples. ResultsWe obtained 11,191 high-quality nuclei comprising neutrophils (53.8%), monocytes (26.1%), lymphocytes (17.8%), and non-immune cells (2.4%). Neutrophils segregated into Nascent, Quiescent, and Interferon-Activated states. Monocytes exhibited classical phenotypes that include interferon-activated states (characterized by expression of VCAN or PROK2) and CXC-chemokine expressing states (characterized by expression of CXCL5 or CXCL8). Lymphocytes were mainly naive and central memory CD4 T cells. Cell-cell signaling analysis predicted strong CXC chemokine signaling from monocytes to neutrophil subsets and IL-1 family-driven inflammatory responses across multiple populations. Type I and III interferon signaling defined a neutrophil population not previously described in the central nervous system. ConclusionThis study delineates the diverse cellular immune landscape of CSF after IVH. Transcriptomic profiles reveal interferon, IL-1, and CXC chemokine signaling networks as potential therapeutic targets to mitigate secondary injury.

genomics↗

bioRxiv2025

RNA localization to nuclear speckles follows splicing logic

Nuclear speckles are membraneless organelles implicated in multiple RNA processing steps. In this work, we systematically characterize the sequence logic determining RNA localization to nuclear speckles. We find extensive similarities between the speckle localization code and the RNA splicing code, even for transcripts that do not undergo splicing. Specifically, speckle localization is enhanced by the presence of unspliced exon-like or intron-like sequence features. We demonstrate that interactions required for early splicesomal complex assembly contribute to speckle localization. We also show that speckle localization of isolated endogenous exons is reduced by disease-associated single nucleotide variants. Finally, we find that speckle localization strongly correlates with splicing kinetics of splicing-competent constructs and is tightly linked to the decision between exon inclusion and skipping. Together, these results suggest a model in which RNA speckle localization is associated with the formation of the early spliceosomal complex and enhances the efficiency of splicing reactions. HighlightsO_LISequences containing hallmarks of pre-mRNA dictate speckle localization C_LIO_LIRNA speckle localization is coupled to early spliceosome assembly C_LIO_LIDisease-associated single nucleotide variants reduce localization of isolated exons C_LIO_LIRNA speckle localization strongly correlates with splicing kinetics C_LI O_FIG O_LINKSMALLFIG WIDTH=191 HEIGHT=200 SRC="FIGDIR/small/656493v1_ufig1.gif" ALT="Figure 1"> View larger version (57K): org.highwire.dtl.DTLVardef@1efbb02org.highwire.dtl.DTLVardef@1deb7aaorg.highwire.dtl.DTLVardef@ec1ec2org.highwire.dtl.DTLVardef@1d78b72_HPS_FORMAT_FIGEXP M_FIG Graphical Abstract C_FIG

biochemistry↗

bioRxiv2025

Peripheral blood cell-type and sex-specific signatures of alcohol misuse revealed by single-cell transcriptomics

Alcohol use disorder (AUD) is a complex condition with diverse molecular underpinnings. Chronic alcohol exposure is associated with alterations in both innate and adaptive immune cell populations that in turn contribute to AUD susceptibility and severity, suggesting a bidirectional gene-environment relationship. Yet, most clinical AUD studies of this system have relied on indirect measurement of immune function and/or assessment of only a small number of cell types or immune markers, approaches which cannot capture the immune systems inherent complexity and cellular heterogeneity. Therefore, in order to better characterize immune dysregulation in AUD, the goal of this study was to use single cell RNA sequencing (scRNA-seq) in peripheral blood mononuclear cells (PBMCs) to compare immune cell proportions and differential gene expression between individuals with AUD and healthy controls. Our findings highlight a distinct disproportionality of lymphocytes and monocytes in individuals with AUD and healthy controls as well as sex and dose-dependent alterations immune cell functional characteristics. These peripheral blood cell-type proportions and dose-dependent signatures of alcohol exposure may aid in developing more targeted and effective pharmacological interventions for AUD. Results also highlight the importance in including sex as a biological variable in AUD research, particularly when examining immune function.

genomics↗

bioRxiv2025

Drosophila Metaxin-2 controls beta-barrel biogenesis and muscle growth in a stage-dependent fashion

Metaxin-2 (Mtx2) is an evolutionarily conserved mitochondrial outer membrane protein. Mutations in human Mtx2 cause mandibuloacral dysplasia (MADaM), a progeroid disorder. However, the pathologic mechanisms of Mtx2 loss-of-function remain largely unknown. Using Drosophila, we show Mtx2 null mutants exhibit pupal lethality, rescued by Drosophila or human Mtx2, underscoring functional conservation across species. Tissue-specific conditional knockout and rescue experiments reveal muscle as a critical site of dMtx2 action, with alternations of myofibril assembly and myogenic proteins being observed in dMtx2 mutants. Structural and functional mitochondrial abnormalities are also detected, verifying dMtx2s function in mitochondrial homeostasis. Notably, Mtx2 deficiency affects beta-barrel protein biogenesis and muscle development in pupa but not in larva, demonstrating Mtx2s dynamic regulation in mitochondrial proteostasis and muscle development. Our results elucidate mitochondrial mechanisms driving potential muscle-autonomous defects in MADaM patients and highlights stage-specific Mtx2 function as a prospective therapeutic target for this progeroid syndrome.

developmental biology↗

bioRxiv2025

OPTRACE: Optical Imaging Guided Transplantation and Tracking of Cells in the Mouse Brain

Intracerebral cell transplantation holds great promise for treating stroke and other neurological disorders, but progress is hindered by challenges in precise cell delivery and limited methods for monitoring post-engraftment dynamics. To address these limitations, we introduce OPTRACE (OPtical imaging-guided Transplantation and tRAcking of CElls), a two-step framework that integrates precision delivery with longitudinal in vivo tracking to advance cell therapy. During transplantation, OPTRACE enables real-time visualization using cost-effective translucent glass micropipettes, guided by predictive modeling to optimize delivery parameters. A novel pulse-elevation injection technique further enhances the precision of superficial cortical engraftment. Following transplantation, multicolor labeling combined with two-photon fluorescence microscopy permits longitudinal single-cell tracking, revealing host microglial responses, and altered neuronal calcium signaling at the graft interface. By unifying high-resolution delivery and tracking strategies, OPTRACE provides a powerful platform for elucidating engraftment biology and accelerating the development of effective cell-based interventions for neurological diseases.

bioengineering↗

bioRxiv2025

Perivascular matrix densification dysregulates angiogenesis and activates pro-inflammatory endothelial cells

Fibrosis is central to numerous fatal conditions including solid cancers, pulmonary fibrosis, cirrhosis and post-infarct cardiac fibrosis amongst many others, thereby collectively contributing to 45% of all deaths in developed nations. The potential for fibrosis across most organ systems may stem from its connections to wound healing and the ubiquitous presence of vascular endothelium. Endothelial cells (ECs) and angiogenesis, cells and associated biological program central to wound healing, have been heavily implicated in many organ-specific fibroses, but the relationship between angiogenesis and fibrogenesis remains debated and little has been established in terms of how the EC phenotype governs tissue healing vs. fibrosis. Here, we examine a murine lung injury model enabling EC lineage tracing and observe the invasion of aberrant ECs from the bronchial microvasculature following lung injury along with concurrent densification of matrix fibers surrounding these vessels. To investigate the underlying mechanisms governing their appearance, we established a microphysiological system (MPS) of arteriole/venule-scale microvessels embedded within a tunable stromal mimetic matrix and find that heightened extracellular matrix fiber density activates ECs, drives endothelial to mesenchymal transition, and promotes aberrant tip EC (ATEC) invasion into the matrix. ATECs remain adherent to fibrotic matrix and possess a pro-inflammatory phenotype that secretes TGF-{beta}2. Notably, our studies establish that the formation of ATECs is gated by destabilization of endothelial adherens junction upon EC adhesion to fibrous matrix, and associated regulation of TGF-{beta} signaling that is mediated by a novel VE-cadherin - TGF-{beta}R2 axis. The current lack of effective anti-fibrotic therapies suggests potential critical involvement of other cell types such as ECs, and our findings suggest new contributions of ECs to fibrotic progression that may better inform future targets for novel anti-fibrotic therapeutics.

bioengineering↗

bioRxiv2025

The representation of facial emotion expands from sensory to prefrontal cortex with development

Facial expression recognition develops rapidly during infancy and improves from childhood to adulthood. As a critical component of social communication, this skill enables individuals to interpret others emotions and intentions. However, the brain mechanisms driving the development of this skill remain largely unclear due to the difficulty of obtaining data with both high spatial and temporal resolution from young children. By analyzing intracranial EEG data collected from childhood (5-10 years old) and post-childhood groups (13-55 years old), we find differential involvement of high-level brain area in processing facial expression information. For the post-childhood group, both the posterior superior temporal cortex (pSTC) and the dorsolateral prefrontal cortex (DLPFC) encode facial emotion features from a high-dimensional space. However, in children, the facial expression information is only significantly represented in the pSTC, not in the DLPFC. Further, the encoding of complex emotions in pSTC is shown to increase with age. Taken together, young children rely more on low-level sensory area than on the prefrontal cortex for facial emotion processing, suggesting that the prefrontal cortex matures with development to enable a full understanding of facial emotions, especially complex emotions that require social and life experience to comprehend.

neuroscience↗

bioRxiv2025

Dominant spinal muscular atrophy linked mutations in the cargo binding domain of BICD2 result in altered interactomes and dynein hyperactivity

Cytoplasmic dynein-1 (dynein) is responsible for the transport of most cellular cargo towards the minus end of microtubules. Dynein activation requires the multi-subunit dynactin complex and an activating cargo adaptor. The adaptors serve to link dynein with cargo and to fully activate the motor. Mutations in one of these activating adaptors, Bicaudal-D2 (BICD2), are associated with a neurodegenerative disease called Spinal Muscular Atrophy with Lower Extremity Predominance (SMALED2). The molecular defect that underlies SMALED2 is largely unknown. In addition to interacting with dynein, BICD2 has also been shown to associate with KIF5B, a plus-end directed microtubule motor. We hypothesized that interactome changes associated with mutant versions of BICD2, and the resulting differences in cargo transport, might underlie the etiology of SMALED2. To test our hypothesis, we first defined the interactome of wild-type BICD2. This led to the identification of known BICD2 interacting proteins in addition to potentially novel cargo such as components of the HOPS complex, a six-subunit complex involved in endo-lysosomal trafficking. We next determined the interactome of three SMALED2 linked mutants in BICD2, two of which reside in the cargo binding domain. Interestingly, all three mutations resulted in BICD2-mediated dynein hyper-activation. Furthermore, all three mutants were associated with interactome changes. One of these mutants, BICD2_R747C, was deficient in binding to HOPS complex components and the nucleoporin RANBP2. In addition, this mutant also resulted in a gain of function interaction with GRAMD1A, a protein localized to the endoplasmic reticulum. This gain of function interaction resulted in mis-localization of GRAMD1A in BICD2_R747C expressing cells. Collectively, our results suggest that dynein hyperactivity, interactome changes, and cargo transport defects might contribute to the symptoms associated with SMALED2.

cell biology↗

bioRxiv2025

SEHI-PPI: An End-to-End Sampling-Enhanced Human-Influenza Protein-Protein Interaction Prediction Framework with Double-View Learning

Influenza continues to pose significant global health threats, hijacking host cellular machinery through protein-protein interactions (PPIs), which are fundamental to viral entry, replication, immune evasion, and transmission. Yet, our understanding of these host-virus PPIs remains incomplete due to the vast diversity of viral proteins, their rapid mutation rates, and the limited availability of experimentally validated interaction data. Additionally, existing computational methods often struggle due to the limited availability of high-quality samples and their inability to effectively model the complex nature of host-virus interactions. To address these challenges, we present SEHI-PPI, an end-to-end framework for human-influenza PPI prediction. SEHI-PPI integrates a double-view deep learning architecture that captures both global and local sequence features, coupled with a novel adaptive negative sampling strategy to generate reliable and high-quality negative samples. Our method outperforms multiple benchmarks, including state-of-the-art large language models, achieving a superior performance in sensitivity (0.986) and AUROC (0.987). Notably, in a stringent test involving entirely unseen human and influenza protein families, SEHI-PPI maintains strong performance with an AUROC of 0.837. The model also demonstrates high generalizability across other human-virus PPI datasets, with an average sensitivity of 0.929 and AUROC of 0.928. Furthermore, AlphaFold3-guided case studies reveal that viral proteins predicted to target the same human protein cluster together structurally and functionally, underscoring the biological relevance of our predictions. These discoveries demonstrate the reliability of our SEHI-PPI framework in uncovering biologically meaningful host-virus interactions and potential therapeutic targets.

bioinformatics↗

bioRxiv2025

High-throughput single-cell DNA methylation and chromatin accessibility co-profiling with SpliCOOL-seq

DNA methylation and chromatin accessibility are fundamental epigenetic mechanisms that orchestrate gene expression programs, define cellular states, and drive developmental trajectories. scCOOL-seq has enabled simultaneously measuring the two modalities in the same single cells, but in quite a low throughput manner. We present single-cell split-pool ligation-based multi-omics sequencing technology (SpliCOOL-seq), which improves the throughput to thousands of cells by combining split-pool ligation based single-cell indexing after in situ tagmentation with universal Tn5 transposase and scCOOL-seq. SpliCOOL-seq achieved higher sensitivity than previous high throughput single-cell DNA methylation sequencing methods and can clearly distinguish different lung cancer cells based on both genetic and multiple epigenetic modalities. We show that the two DNMT inhibitors, 5-Azacitidine and Decitabine, both cause large scale demethylation but in distinct patterns. Applied to the primary lung tumor, SpliCOOL-seq clearly captured subclones within the tumor lesion and revealed candidate genes related to tumorigenesis. Furthermore, we presented the first report on the heterogeneity of scDNAm age acceleration among tumor subclones as predicted from a single-cell perspective. In conclusion, SpliCOOL-seq achieves parallel profiling of whole genome DNA methylation and chromatin accessibility in the same individual cells in a high-throughput manner and is hopefully used to illustrate regulatory interactions under different cell states.

genetics↗

bioRxiv2025

Optimizing Metagenome Analysis for Early Detection of Colorectal Cancer: Benchmarking Bioinformatics Approaches and Advancing Cross-Cohort Prediction

Colorectal cancer (CRC) continues to be a major global public health challenge. Extensive research has underscored the critical role of the gut microbiome for diagnostics of CRC. However, early-stage prediction of CRC, particularly at the precancerous adenomas (ADA) stage, remains challenging due to the instability of microbial features across cohorts. In this study, we conducted a systematic analysis of 2,053 gut metagenomes from 14 globally-sampled public cohorts and a newly recruited cohort. Despite substantial regional and cohort-level heterogeneity in microbiome composition, we elucidated that the consistent dynamic patterns of microbial signatures are the fundamental for CRC detection. These patterns enabled robust performance in both inter-cohort and independent validations using an optimized bioinformatics framework. In contrast, such basis was lacking in ADA-associated microbial markers, limiting the generalizability of early detection models. To address this, we developed an instance-based transfer learning approach, Meta-iTL, which effectively leveraged knowledge from existing datasets to detect CRC risk at the ADA stage in the newly recruited cohort. Thus, Meta-iTL overcomes challenges posed by cohort-specific variability and limited data availability, advances the application of non-invasive approaches for the early screening and prevention of CRC.

bioinformatics↗

bioRxiv2025

ZEB2 signaling is essential for ureteral smooth muscle cell differentiation and maintenance

Mowat-Wilson Syndrome (MWS) is a multiple congenital anomaly syndrome caused by mutations in the ZEB2, which plays a critical role in cell fate determination and differentiation during development. Congenital anomalies of the kidney and urinary tract (CAKUT) have been reported in MWS patients. However, the role of ZEB2 in urinary tract development and the cellular and molecular mechanism underlining the CAKUT phenotypes in MWS remains unknown. We performed ZEB2 protein expression analysis in the developing mouse ureter. We generated Zeb2 ureteral mesenchyme-specific conditional knockout mice by crossing Zeb2 floxed mice with Tbx18Cre mice (Zeb2 cKO) and analyzed the urinary tract phenotypes in Zeb2 cKO mice and wild-type littermate controls by gross and histological examination. Ureteral cellular and molecular phenotypes were studied using TAGLN, ACTA2, FOXD1, POSTN, CDH1, TBX18, and SOX9 ureteral cell-specific markers. We found that ZEB2 is expressed in TBX18+ ureteral mesenchymal cells during mouse ureter development. Deletion of Zeb2 in developing ureteral mesenchymal cells causes hydroureter and hydronephrosis phenotypes, leading to obstructive uropathy, kidney failure, and early mortality. Cellular and molecular marker analyses showed that the TAGLN+ACTA2+ ureteral smooth muscle cells (SMCs) layer is not formed in Zeb2 cKO mice at E15.5, but the FOXD1+ and POSTN+ tunica adventitia cells layer is significantly expanded compared to wild-type controls. CDH1+ urothelium cells are reduced considerably in the Zeb2 cKO ureters at E15.5. Mechanistically, we found that Zeb2 cKO mice have significantly decreased TBX18 expression but an increased SOX9 expression in the developing ureter at E14.5 and E15.5 compared to wild-type littermate controls. Our results show that ZEB2 is essential for ureter development by maintaining ureteral mesenchymal cell differentiation into normal ureteral SMCs. Our study also shed new light on the pathological mechanism underlying the developmental abnormalities of the urinary tract phenotypes in MWS patients. Author SummaryHydroureter and hydronephrosis are common congenital anomalies with high incidence (1:100 to 1:500) that can lead to obstructive uropathy and renal failure in the pediatric population. Mowat-Wilson Syndrome (MWS), caused by heterozygous mutations in the ZEB2 gene, is a genetic disease with multiple congenital developmental defects, including hydroureter and hydronephrosis. However, the molecular function of ZEB2 in ureter development and pathogenesis of hydroureter and hydronephrosis remains unknown. Here, we show that ZEB2 is expressed in the developing ureteral mesenchymal cells, and deletion of ZEB2 in ureteral mesenchymal cells leads to ureteral smooth muscle cell loss, which is replaced by tunica adventitia cells causing hydroureter and hydronephrosis phenotype. Hence, our work not only demonstrates the critical role of ZEB2 in ureter development but also provides molecular mechanisms of urinary tract anomalies in MWS patients.

developmental biology↗

bioRxiv2025

Graph-based Contrastive Learning Enables Unified Integration and Niche Transfer Across Single-Cell and Spatial Multi-Omics

The rapid growth of single-cell and spatial omics has outpaced computational methods capable of unifying these data into a cohesive framework for tissue atlas construction and cross-sample analysis. A critical bottleneck lies in the inability of existing tools to co-embed cells from diverse technologies--spanning transcriptomics, epigenomics, and proteomics--into a shared reference space while preserving spatial architecture and molecular specificity. Here, we present Garfield (Graph-based Contrastive Learning Enables Fast Single-Cell Embedding), a geometric deep-learning framework that addresses these challenges through spatially or molecularly aware cell embedding. Leveraging a graph contrastive learning framework, Garfield learns a shared embedding space for data generated by diverse technologies, enabling seamless construction and querying of spatial reference atlases. Our results show that Garfield consistently outperforms state-of-the-art benchmark models in identifying spatial niches across multiple datasets. We further demonstrate Garfields versatility by applying it to multi-modal spatial data, including gene expression and chromatin accessibility, where it successfully identifies distinct niches in the mouse brain. Notably, Garfield reveals tumor microenvironment heterogeneity in non-small cell lung cancer and breast cancer, uncovered conserved, barrier-like immune niches at tumor margins orchestrating CD80-mediated T cell-B cell-dendritic cell interactions and IFN-/B cell activation pathways, forming spatially coordinated immune surveillance hubs. These findings underscore Garfields potential to advance spatial omics research by offering a robust, scalable solution for integrating and interpreting complex spatial data across diverse tissue types and modalities.

bioinformatics↗